Abstract: Objective To explore the influencing factors of new fractures after percutaneous vertebroplasty (PVP) in the patients with osteoporotic thoracolumbar fractures (OTF), and to establish a nomogram early warning model and verify the predictive performance of the model. Methods A total of 193 patients with OTF who underwent PVP in our hospital from January 2017 to December 2018 were retrospectively selected as the modeling group, and 63 patients with OTF who underwent PVP from January to December 2019 were selected as the validation group. The end point of follow-up was 2 years after surgery or the time of diagnosis of new postoperative fractures. The incidence of postoperative fractures in the modeling group was observed, and the independent risk factors for new postoperative fractures were screened by univariate and multivariate Cox proportional hazards models. R software was used to establish a nomogram model to predict the risk of new fractures after surgery, and the predictive effect of the model was verified internally and externally. Results The incidence rate of new fractures after PVP was 21.76% (42/193). Multivariate Cox regression analysis showed that lumbar spine T value, intravertebral fissure-like degeneration, height recovery rate of injured vertebral anterior edge, bone cement injection volume, and bone cement leakage were the independent risk factors for new fractures after PVP in the patients with OTF. Based on the multivariate Cox regression analysis results, a nomogram model for predicting new postoperative fractures was established. The C-index of the prediction model was 0.730 (95%CI: 0.706-0.784) in the validation group and 0.796 (95%CI: 0.724-0.836) in the modeling group, and the areas under the receiver operating characteristic curve for predicting the incidence of new fractures at 1 year and 2 years after surgery were 0.778, 0.721 in the validation group and 0.820, 0.772 in the modeling group, respectively. The slopes of the calibration curves of the two groups were both close to 1, and the threshold probability of the decision curve analysis (DCA) was 0.05-1, and the model showed a positive net benefit. Conclusions This study establishes a nomogram prediction model based on five independent risk factors of new fractures after PVP for OTF, which has a good predictive performance and can provide a certain reference value for clinical medical staff to prevent fractures after PVP.

Key words: osteoporosis, thoracolumbar fractures, percutaneous vertebroplasty, new vertebral fracture